Re: [Roll] Review request for draft-clausen-lln-rpl-experiences-11

[Philip Levis <[e-mail]>]

> On May 16, 2018, at 2:31 PM, Ines Robles <[e-mail]> wrote:
> 
> Dear all,
> 
> This  Individual Submission depicts observations of RPL.
> 
> https://tools.ietf.org/html/draft-clausen-lln-rpl-experiences-11 <https://tools.ietf.org/html/draft-clausen-lln-rpl-experiences-11>
> 
> Please let us know your opinion.
> 
> Thanks,
> 
> Ines and Peter

Your observations about Trickle are correct — its transmission rate can asymptotically approach zero in a stable network, but dynamic networks that trigger its reset conditions cause it to transmit more frequently. E.g., 

https://sing.stanford.edu/site/publications/ctp-tosn14.pdf <https://sing.stanford.edu/site/publications/ctp-tosn14.pdf>

shows experimental results from CTP (which motivated much of the design behind upward routes and the use of Trickle). Figures 35, 36 and 37 show how nodes in dynamic conditions (people moving around) send more control packets. These results are in contrast to Figures 15 and 17(b), which show how in stable conditions control traffic asymptotically approaches some low rate. 

However, it’s not clear to me what else you can do: if your physical connectivity is changing significantly and frequently, you need to exchange control messages more frequently to maintain the routing topology. The basic premise of Trickle is that you adapt the rate to what the topology needs: rather than transmit at a fixed rate, which is either too slow (during high dynamics) or too fast (when the topology is stable), the protocol reactively repairs the topology.

Phil